A method to accurately determine SNP allelic frequencies from pooled DNA samples would greatly increase the throughput and, therefore, reduce the cost of SNP frequency determination. Such a method would also provide a reliable approach for selecting appropriate SNP's from databases for study. For example, databases of human SNPs comprise millions of SNP sites. Choosing low frequencies SNPs for a disease association study would require that a larger number of individual organisms be genotyped in order to obtain the statistically meaningful data. It has been suggested that SNPs for such studies should have a minimum allelic frequency of at least 10%. Techniques for allelic frequency determination are known.